How Evidence Works

Last updated: August 28, 2026.

Findings are evidence-backed, not invented

Every finding ShipForge reports is tied to a concrete location in your source code: a file path, a line or region, and a short snippet of the offending code. A finding without evidence is not reported. If evidence is insufficient to support a claim, ShipForge labels the finding “Could not verify” rather than confirming a defect — it does not assert the feature is missing, only that it could not be confirmed. Findings ShipForge can confirm from concrete evidence are labeled “Deterministically verified”; reasoned improvement suggestions are labeled “AI recommendation” and are not confirmed defects.

Severity reflects real launch risk

Each finding carries a severity — critical, high, medium, or low — reflecting the risk it poses to a successful app launch or store review. Severity is assigned based on the evidence and the category of issue, not on a guess about your business.

The difference between automated and verified

The free, live tool today is the Launch Score — a 15-question self-assessment you score yourself, with no backend scan. A fully automated Launch Doctor (self-serve upload, scan, and report) is on the roadmap and is not live yet. The Launch Doctor Review ($99 Founding Beta) is an evidence-backed automated review: deterministic checks read your code, an AI explains and prioritizes the findings, an independent model challenges the reasoning, and automated quality gates check high-stakes items before delivery — a paid, quality-controlled product with limited availability. ShipForge is not currently able to offer a qualified human technical reviewer.

What we never claim

ShipForge does not claim that passing its checks guarantees App Store or Google Play approval, a particular search ranking, or any revenue outcome. Store approval is decided by Apple and Google. ShipForge helps you find and fix issues before submission — it does not approve apps on their behalf.

Score methodology

Your Launch Score is a 0–100 readiness indicator derived from your answers to 15 self-assessment questions across categories (security, store-readiness, configuration, monetization, and launch infrastructure). It is a directional signal based on your own self-ratings, not a grade or guarantee, and not a scan of your code. A higher score means fewer self-reported launch risks.